Description of problem: 1. Publish VM to a template 2. Run provider full refresh (to bypass bug 1517852). 3. VM provision from the above template: Select the template, choose cluster/datastore, Select network, Hardware - no changes done here -> It fails with error in UI, regarding the Hardware tab: "Memory Limit is supported for RHV 4.1 and above. Current provider version is ." This is though the version of the RHV provider connected is RHV-4.1.7. Version-Release number of selected component (if applicable): CFME-5.9.0.10+Graph refresh ON/RHV-4.1.7 Additional info: * Note that in the error message it seems the RHV version is unknown. * On CFME-5.9.0.10 without Graph refresh the VM provision succeeds. * The CFME was connected to 2 small RHV providers, and had 8G memory. On RHV UI, where this CFME VM runs, it's memory reach 95% - That might be related to Graph refresh.
Created attachment 1359533 [details] vm_provision_max_memory_failure.png
Created attachment 1359534 [details] evm.log see template named temp_name
Created attachment 1359535 [details] automation.log
Ilanit - does it work without graph refresh?
Apparently, the missing RHV version is a known issue on CFME Graph refresh, and is already being handled. Adam, Would you please point to the relevant PR/BZ? Thanks.
It will be handled by us.
I attempted to reproduce it following the steps specified in comment #1 without luck. Please retest with the latest code or specify additional steps that need to performed to reproduce it.
Piotr did you run this with only graph refresh on? I thought we knew the api_version wasn't set by graph refresh.
Adam, I used regular refresh, changed the config setting and restarted manageiq. I will check with using graph refresh from the beginning.
Yes, when using only graph refresh I was able to reproduce the issue.
Verified on CFME-5.10.0.15 (Graph refresh is ON by default)/RHV-4.2.3 VM provision from template end up successfully.